Accurate prediction of free-surface flows is a critical challenge in hydraulic and environmental engineering, governing flood risk assessment, debris flows, or sediment transport. While the standard Shallow Water Equations (SWE) are computationally efficient, they fail to capture vertical flow structures, whereas full 3D Navier-Stokes simulations remain prohibitively expensive for large-scale applications. A compromise lies in moment-based approximations, which extend the two-dimensional structure of the SWE by capturing vertical velocity profiles or non-hydrostatic contributions. These models are hierarchical, enabling variable fidelity through an adjustable polynomial basis. However, their adoption in engineering software remains limited; existing implementations typically support only specific models with fixed polynomial degrees, thereby negating the advantage of adaptive fidelity. Here, we present Zoomy, a framework for simulating hierarchical models, such as the Shallow Moment Equations (SME) and Vertically Averaged and Moment Equations (VAM), for engineering-relevant flows. We show that by leveraging Zoomy’s automated kernel generation, flexible-fidelity models can be rapidly deployed to capture complex vertical phenomena—such as secondary currents—that are invisible to standard shallow-water models. This approach demonstrates that hierarchical moment equations can be used with the same ease as standard models, providing a scalable pathway to 3D-like accuracy with computational efficiency comparable to that of 2D solvers. Ultimately, this framework empowers researchers to focus on the physical properties of free-surface flow models rather than the complexities of their numerical implementation.